abliterated models, refusals nuked
🏗️ Building on HF
josephmayo
josephmayo
AI & ML interests
why dont we make smart models that can fit on consumer hardware
Recent Activity
liked a model 1 day ago
Qwen/Qwen2.5-Coder-7B-Instruct liked a dataset 5 days ago
josephmayo/curated-openbmb-code-math reacted to sergiopaniego's post with 👍 7 days ago
Frontier models use distillation as a step of their post-training pipelines.
In 2026 it has three jobs: compress a big model into a small one, merge RL experts into a single model, and let a model teach itself.
I wrote up which frontier models use each one and how: https://huggingface.co/blog/sergiopaniego/distillation-2026
It pairs with Class 2 of the Training an Agent series Ben and I are doing, where we teach these techniques hands-on with TRL! Organizations
None yet
models 25
josephmayo/Holo-3.1-9B-Coder
Text Generation • 9B • Updated • 3.24k • 3
josephmayo/ZAYA1-8B-Coder-GGUF
9B • Updated • 2.4k
josephmayo/ZAYA1-8B-Coder
Text Generation • 9B • Updated • 37
josephmayo/ZAYA1-8B-Coder-LoRA
Updated • 9 • 1
josephmayo/gemma-4-E4B-it-Coder-GGUF
8B • Updated • 382
josephmayo/gemma-4-E4B-it-Coder
Image-Text-to-Text • 8B • Updated • 43
josephmayo/Holo-3.1-4B-Coder-GGUF
Text Generation • 4B • Updated • 227
josephmayo/Holo-3.1-4B-Coder
Text Generation • 4B • Updated • 29 • 1
josephmayo/Holo-3.1-4B-Coder-LoRA
Text Generation • Updated • 12
josephmayo/LFM2.5-8B-A1B-Coder-GGUF
8B • Updated • 396 • 1